AI Remediation Engine
Hookmetry AI is an optional escalation after deterministic forensics. It converts an eligible failure context into one root-cause explanation and copy-paste code fixes for common application stacks.
AI analysis is user-triggered from the dashboard, the REST API, or the CLI. It is never required for webhook ingestion, and an AI provider failure does not invalidate the captured log.
Before you begin
- A captured webhook log with a validation or destination failure.
- An available AI allowance for the account plan.
- Hookmetry AI available for the account.
What is sent for analysis
- Provider name, error code, failed reconstruction step, and structural fingerprint fields such as lengths, encodings, and boolean indicators.
- For destination failures, the HTTP status and transport error context may be used when available.
- Raw webhook payloads and configured secrets are intentionally excluded from the validation-failure prompt.
Output
- A plain-English explanation of the likely root cause.
- Language-specific snippets for Node.js, Python, Java/Spring Boot, and Go when the model returns them.
- The provider, error code, generation time, and Hookmetry AI model label for traceability.
Plan allowances
| Plan | AI allowance |
|---|---|
| Free | 3 lifetime requests. |
| Pro | 100 requests using the current account counter. |
| Team | Unlimited. |
| Business | Unlimited. |
| Enterprise | Defined by contract. |
Caching and failure behavior
Results are cached by provider, error code, and structural fingerprint for seven days. A matching cached diagnosis can be written to another log without another model call. If the AI key is unavailable, the model response is invalid, or the call fails, the request returns an AI failure while the deterministic webhook diagnosis remains available.